This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
During a 2025 hiring surge, U.S. Immigration and Customs Enforcement (ICE) utilized an AI résumé screening tool that incorrectly identified applicants as having law enforcement experience based on the keyword 'officer'. This error caused recruits without proper policing backgrounds to be assigned to a shortened training pathway instead of the required in-person academy course. ICE identified the issue mid-fall, conducted manual reviews, and reassigned affected recruits for additional training.
U.S. Immigration and Customs Enforcement (ICE) reportedly used an AI-assisted résumé screening tool during a 2025 hiring surge that misclassified some applicants as having law-enforcement experience. As a result, certain recruits without policing backgrounds were allegedly routed into a shortened training pathway. ICE reportedly identified the error, reviewed résumés manually, and reassigned affected recruits for additional training.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
The AI system failed to perform reliably by using overly simplistic keyword matching (flagging any instance of the word 'officer'), demonstrating a lack of capability in semantic resume parsing.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by the AI resume screening tool failing to accurately categorize applicants, which was an unexpected and unintended outcome of its deployment.
An ICE AI resume screening tool misclassified approximately 200 recruits due to simple keyword matching, sending them to a shortened training program. The error was caught and corrected before any recruits were deployed to active duty, resulting in minor disruption to agency operations and highlighting risks of overreliance on automated hiring tools.
The AI system functioned as a tool to scan and categorize resumes based on keywords like 'officer'. It performed a specific, well-defined function on direct inputs without autonomous planning or interactive assistance.
High Risk: The system is an employment and worker management system used in recruitment ('using AI to scan resumes'), which is explicitly classified as High Risk under the EU AI Act.
The AI tool was deployed to scan applicant resumes and categorize/group them based on whether they had prior law enforcement experience to determine their training pathway.
The reports provide consistent details about the AI tool's specific failure mode (keyword matching on 'officer'), the number of affected individuals (about 200), and the corrective actions taken by ICE. The role of the AI is explicitly stated, and there are no major conflicting accounts regarding the AI error itself.
None. The failure is clearly attributed to a simplistic keyword-matching logic in the AI resume screening tool.
Technology
Keyword-Based AI Classification
The tool classified applicants based solely on keyword matching like 'officer'.
Lack of Contextual Understanding
AI failed to distinguish 'compliance officer' from actual law enforcement.
Data Inputs
Unstructured Resume Data
Resume text variations led to keyword mismatches by the screening tool.
Human Factors
Overreliance on AI Screening
Staff relied on AI categorization without immediate manual verification.
Process and Methods
Absence of Initial Manual Review
No manual checks were performed on AI classifications during the hiring surge.
Inadequate AI Validation Process
The AI tool was deployed without rigorous validation of its sorting logic.
Management
Aggressive Hiring Mandate
Pressure to hire 10,000 officers led to rushed deployment of the AI tool.
Inadequate Risk Assessment
Management failed to assess risks of automated resume sorting errors.
